Implementation of RANSAC Algorithm for Feature-Based Image Registration

Abstract

This paper describes the hardware implementation of the RANdom Sample Consensus (RANSAC) algorithm for featured-based image registration applications. The Multiple-Input Signature Register (MISR) and the index register are used to achieve the random sampling effect. The systolic array architecture is adopted to implement the forward elimination step in the Gaussian elimination. The computational complexity in the forward elimination is reduced by sharing the coefficient matrix. As a result, the area of the hardware cost is reduced by more than 50%. The proposed architecture is realized using Verilog and achieves real-time calculation on 30 fps 1024 * 1024 video stream on 100 MHz clock.